impact of data on simulation: from early practices to...
Post on 14-Jul-2020
3 Views
Preview:
TRANSCRIPT
EUROSIM 2001© Tuncer Ören 2001-06-26 1 / 43-
EUROSIM 2001: Shaping Future with Simulation
Impact of Data on Simulation:From Early Practices to Federated
and Agent-Directed Simulation
Delft, June 26-29 2001
Dr. Tuncer ÖrenProfessor Emeritus, University of Ottawa, Canada
oren@site.uottawa.cahttp://www.site.uottawa.ca/~oren/
EUROSIM 2001© Tuncer Ören 2001-06-26 2 / 43-
Aims:To elaborate on
- the types and importance of
- data
- simulation
- the impact of data on simulation
EUROSIM 2001© Tuncer Ören 2001-06-26 3 / 43-
Plan:• Impact of Data: A Milestone Example• Basic Concepts: Belief, fact, data, information,
and knowledge• Where Data Matters in Simulation• 3 Perceptions of simulation• Simulation and Real System: Concurrency• Possibilities for Augmented Reality• Levels of Perception of Simulation• Some Advanced Types of Simulation• Unity in Diversity
EUROSIM 2001© Tuncer Ören 2001-06-26 4 / 43-
Impact of Data: A Milestone ExampleFrom earth-centric universe to sun-centric solar system
C. Ptolemy (100-175) earth-centric view
N. Copernicus (1473-1543) formulation of the sun-centric view
Tycho Brahe (1548-1601) relentless observations (i.e., data)
of the planetary system
J. Kepler (1571-1630) abstraction & formulation
G. Galilei (1564-1642) own observations (i.e., data) and
support of and fight for
Copernican view
EUROSIM 2001© Tuncer Ören 2001-06-26 5 / 43-
Basic Concepts: Belief, fact, data, information, and knowledge (1/4)
Belief is a hypothesis about some unobservable situation.
Beliefs do not need to be true!
Fact (1539) is what makes a belief true or false.
(from factum – a thing done)
EUROSIM 2001© Tuncer Ören 2001-06-26 6 / 43-
Basic Concepts: Belief, fact, data, information, and knowledge (2/4)
Data (1646) plural of datum;
Data means factual information given or admitted, as measurement or statistics, to be used as a basis for reasoning, inferencing, discussion, or calculation.
EUROSIM 2001© Tuncer Ören 2001-06-26 7 / 43-
Basic Concepts: Belief, fact, data, information, and knowledge (3/4)
Knowledge
- (Russell): a sub-class of true beliefs.
- (Minsky): justifiably true beliefs.
- (Hayes-Roth): facts, beliefs, and heuristic rules.
EUROSIM 2001© Tuncer Ören 2001-06-26 8 / 43-
Contradistictions: Belief, fact, data, information, and knowledge (4/4)
(Wildberger)
Data are facts.
Information is data organized for some human purpose.
Knowledge is information and how to use it.
Deciding is acting on information.
EUROSIM 2001© Tuncer Ören 2001-06-26 9 / 43-
• available data• background data• biased data• certain data• certified data• complementary data• conditional data• confidential data• contradictory data• conventional data• corrupted data• customizable data• discrete data ...
(about 150) Types of data:
• abnormal data• abstract data• actual data• affected data• alphabetic data• alphanumeric data• altered data• ambiguous data• analog data• analog input data
EUROSIM 2001© Tuncer Ören 2001-06-26 10 / 43-
(about 150) Types of data:
• divided data• domain-dependent data• domain-independent data• dynamic data• dynamical data• dynamically-changing data• electronic data• empirical data• emulated data• encripted data
• endogenous data• evidential data• exogenous data• external data• externally defined data• externally described data• externally generated data• factual data• formatted data• global data• heterogenously stored data• hidden data• hierarchical data ...
EUROSIM 2001© Tuncer Ören 2001-06-26 11 / 43-
Where Data Matters in Simulation:
To have meaningful and credible results from a simulation study, one must have relevant and correct data.
datasets need to be reliable, validatable, auditable, and replaceable.
Data may be needed in several phases of a simulation study:- to formulate a model (parameter fitting and calibration); - to formulate an environment (static or dynamic)- to validate/verify model and experimental conditions- to generate model behavior (initial conditions, parameters)
EUROSIM 2001© Tuncer Ören 2001-06-26 12 / 43-
3 Perceptions of simulation: (1/11)
“Simulation” is derived from Latin “simulacre”.
“Simulation” has 3 images:
- non-scientific view
- scientific view
- military perception
EUROSIM 2001© Tuncer Ören 2001-06-26 13 / 43-
3 Perceptions of simulation: (2/11)- non-scientific view“Simulation” means fake, a sham object, counterfeit, or imitation. (used since 14th century)
Examples: simulated leather, simulated pearl
There is a confusion of the terms simulation and emulation.
Emulation means: “ambition or endevour to equal or excel others (as in achievement).
Example: a child may emulate her parents; and by doing so she does not simulate them. (However, her behaviour, not being the innate one, may be considered imitation or fake.)
EUROSIM 2001© Tuncer Ören 2001-06-26 14 / 43-
3 Perceptions of simulation: (3/11)- scientific view
“Simulation” is goal-directed experimentation with dynamic models.
When the experimentation cannot or should not be done on the real system, one can perform it using a dynamic model and hence use simulation.
“Simulation” is the contemporary sine qua non technique of Francis Bacon’s (1561-1626) scientific method which is based on experimentation. (as advocated in his Novum Organum published in 1620.)
EUROSIM 2001© Tuncer Ören 2001-06-26 15 / 43-
3 Perceptions of simulation: (4/11)- scientific view
“Until we attempt to simulate a system, we don’t realize how little we know* about it”
Donald Knuth
* both the dynamics and
the relevant parameters (i.e., data)
EUROSIM 2001© Tuncer Ören 2001-06-26 16 / 43-
3 Perceptions of simulation: (5/11)- scientific view
-(from a systemic point of view) simulation can be used to find the values of output, input, or state variablesof a system; provided that the values of the two other types of variables are known. (W. Karplus, 1976).
input variable output variablestate variable
EUROSIM 2001© Tuncer Ören 2001-06-26 17 / 43-
3 Perceptions of simulation: (6/11)- scientific view
input variable output variablestate variable
Type of problem: GivenFind
Analysis Output input state
Design State input output
Control Input state output
EUROSIM 2001© Tuncer Ören 2001-06-26 18 / 43-
(7/11)Input dataSimulation
Digital sensory input dataAnalog sensory input data
Sensor Simulation
Learned data Sensory data
Neural net
Simulation
Inter-simulation dataInput data
Another simulation
Simulation
Types of Simulation Input Data
EUROSIM 2001© Tuncer Ören 2001-06-26 19 / 43-
3 Perceptions of simulation: (8/11)- military perception
Military perception of simulation can be summarized as “All but war is simulation.”
3 types of military simulation:
- Live simulation
- Constructive simulation
- Virtual simulation
EUROSIM 2001© Tuncer Ören 2001-06-26 20 / 43-
3 Perceptions of simulation: (9/11)- military perception: Live simulation
In live simulation, experimentation is performed with simulated (fake) ammunition and real system acting in real environment.
In live simulation, real people and real equipment are both augmented with special sensors to act as target designators.
Live simulation can best be conceived as a special case of augmented reality simulation.
EUROSIM 2001© Tuncer Ören 2001-06-26 21 / 43-
3 Perceptions of simulation: (10/11)- military perception: Constructive simulation
Constructive simulation is war gaming.
Forces, equipment, and environment are represented by models.
At decision points, decision makers inject their decisions to the simulation system.
Constructive simulation fits to the scientific definition of simulation with war gaming connotation.
EUROSIM 2001© Tuncer Ören 2001-06-26 22 / 43-
3 Perceptions of simulation: (11/11)- military perception: Virtual simulation
Virtual simulation is military simulation where virtualequipment –namely, a physical model of the system– is used for training purposes.
In non-military applications the term simulator is used when a physical model of the system is used.
When the physical model has a man-in-the-loop, simulators are used for training purposes.
Virtual simulation fits to the scientific definition of simulation with simulator connotation.
EUROSIM 2001© Tuncer Ören 2001-06-26 23 / 43-
Simulation and Real System: Concurrency (1/9)
2 possibilities:
- Stand-alone simulation
- On-line simulation
EUROSIM 2001© Tuncer Ören 2001-06-26 24 / 43-
Simulation and Real System: Concurrency (2/9)
Stand-alone simulation
Stand-alone simulation is use of simulation independent of the real system.
There are 3 purposes:
- Pure experimentation
- Training to develop skill in the use of hardware
- Training to enhance decision making skill
EUROSIM 2001© Tuncer Ören 2001-06-26 25 / 43-
Simulation and Real System: Concurrency (3/9)
Stand-alone simulation: Pure experimentation
Most common purpose in the use of simulation for both civilian and military applications.
This type of usage supports design, analysis, control, planning, logistic operations, simulation-based acquision, and simulation-based evaluations of products and processes.
EUROSIM 2001© Tuncer Ören 2001-06-26 26 / 43-
Simulation and Real System: Concurrency (4/9)
Stand-alone simulationTraining to develop skill in the use of hardware
A human operator uses a virtual equipment (a simulator) to develop skills to use the equipment.
This usage corresponds
- to simulators (in civilian applications) and
- to virtual simulation (in military applications).
EUROSIM 2001© Tuncer Ören 2001-06-26 27 / 43-
Simulation and Real System: Concurrency (5/9)
Stand-alone simulationTraining to enhance decision making skill
This type of usage is done by gaming simulation.
In civilian applications: business games
In military applications: war games, conflict management simulation, and peace support simulation
EUROSIM 2001© Tuncer Ören 2001-06-26 28 / 43-
Simulation and Real System: Concurrency (6/9)On-line simulation
On-line simulation is use of simulation concurrently with the real system.
There are 3 goals of usages:
- To support the operations of the real system
- To foster on-line diagnosis
- To augment reality
EUROSIM 2001© Tuncer Ören 2001-06-26 29 / 43-
Simulation and Real System: Concurrency (7/9)On-line simulation
To support the operations of the real system
Simulation can provide predictive displays.
EUROSIM 2001© Tuncer Ören 2001-06-26 30 / 43-
Simulation and Real System: Concurrency (8/9)On-line simulation
To foster on-line diagnosis
- Run real system and simulation concurrently
and compare their behaviors.
- A difference may indicate a mulfunction of the real
system.
EUROSIM 2001© Tuncer Ören 2001-06-26 31 / 43-
Simulation and Real System: Concurrency (9/9)On-line simulation
To augment reality
In augmented (or mixed) reality simulation, real and virtual entities (that can be people or equipment) and the environment can exist at the same time.
Hence, operations can take place in a richer augmented reality environment.
Reality is a special case of simulation!
EUROSIM 2001© Tuncer Ören 2001-06-26 32 / 43-
Possibilities for Augmented Reality
Real operator
Virtual operator
Real equipment Virtual equipment
- Live simulation
(a human operator uses virtual guns)
- Simulators
- Virtual simulation
- Automated vehicles(auto pilot, aircraft wo pilot; vehicle wo driver)
e.g., an AI aircraft
EUROSIM 2001© Tuncer Ören 2001-06-26 33 / 43-
Levels of Perception of Simulation: (1/4)
3 levels: simulation as
- a computational activity
- a model-based activity
- a knowledge generation activity
EUROSIM 2001© Tuncer Ören 2001-06-26 34 / 43-
Levels of Perception of Simulation: (2/4)
Simulation as a computational activity
The emphasis is on the generation of model behavior.
(Conventional simulation)
All issues of input data and initialization are applicable.
EUROSIM 2001© Tuncer Ören 2001-06-26 35 / 43-
Levels of Perception of Simulation: (3/4)
Simulation as a model-based activity
In addition to generation of model behavior, computer-aided modelling, model-base management, parameter-base management, and symbolic processing of models are considered.
The role of data in modelling and parameter-base management is primordial.
EUROSIM 2001© Tuncer Ören 2001-06-26 36 / 43-
Levels of Perception of Simulation: (4/4)
Simulation as a knowledge generation activity
The definition of simulation can be interpreted as follows: Simulation is model-based experiential knowledge generation.
This abstraction facilitates the synergy of simulation with other knowledge generation techniques: Optimization, statistical inferencing, reasoning, hypothesis processing.
EUROSIM 2001© Tuncer Ören 2001-06-26 37 / 43-
Some Advanced Types of Simulation: (1/4)
- Federated simulation
- Agent-directed simulation
- Holonic simulation
- Holonic agent simulation
EUROSIM 2001© Tuncer Ören 2001-06-26 38 / 43-
Some Advanced Types of Simulation: (2/4)
Federated simulation
This is an example of interoperability of several simulation studies each called a federate.
Current realization relies on HLA (High Level Architecture).A specific example of the role of data in federated simulation:
Reduction (minimization) of data exchange requirements between federates.
EUROSIM 2001© Tuncer Ören 2001-06-26 39 / 43-
Some Advanced Types of Simulation: (3/4)
Agent-directed simulation
3 possibilities:
- Agent simulation
(simulation of entites represented by agents)
(challenges to quantify motivation and autonomy)
- Agent-based simulation
- Agent-supported simulation
EUROSIM 2001© Tuncer Ören 2001-06-26 40 / 43-
Some Advanced Types of Simulation: (4/4)
Holonic simulation and holonic agent-simulation
Holonic systems are excellent candidates to conceive, model, control, and manage dynamically organizing cooperative systems.
A holonic system is composed of autonomous entities (called holons) that can deliberately reduce their autonomy, when need arise, to collectively achive a goal.
EUROSIM 2001© Tuncer Ören 2001-06-26 41 / 43-
Unity in Diversity
- Art of simulation, Emerging trends*
Methodologies/techniques: Application areas:
A view of the topics of the invited presentations at the EUROSIM 2001 Congress
Impact of data on simulation
* We shall find a way or we shall make one!
EUROSIM 2001© Tuncer Ören 2001-06-26 42 / 43-
Unity in Diversity
A view of the topics of the invited presentations at the EUROSIM 2001 Congress
- Art of simulation, Emerging trends*
Methodologies/techniques: Application areas:
Validation Societal models
Sim environments with adaptive behavior
Atmospheric modelling
Visualization / animation
Integrated water management
Impact of data on simulation
* We shall find a way or we shall make one!
EUROSIM 2001© Tuncer Ören 2001-06-26 43 / 43-
We Have Seen:
• Impact of Data: A Milestone Example• Basic Concepts: Belief, fact, data, information, and
knowledge• Where Data Matters in Simulation• 3 Perceptions of simulation• Simulation and Real System: Concurrency• Possibilities for Augmented Reality• Levels of Perception of Simulation• Some Advanced Types of Simulation• Unity in Diversity
top related